European Journal of Cardio-Thoracic Surgery Advance Access published May 7, 2014
Influence of intraoperative cerebral oximetry monitoring on neurocognitive function after coronary artery bypass surgery: a randomized, prospective study†‡ Zeljko Colaka,*, Marko Borojevicb, Anamarija Bogovicc, Visnja Ivancana, Bojan Biocinab and Visnja Majeric-Koglera a b c
Department of Anesthesiology, Reanimatology and Intensive Care Medicine, University Hospital Center Zagreb, Zagreb, Croatia Department of Cardiac Surgery, University Hospital Center Zagreb, Zagreb, Croatia Psychiatric Hospital Sv. Ivan, Zagreb, Croatia
* Corresponding author. Department of Anesthesiology, Reanimatology and Intensive Care Medicine, University Hospital Center Zagreb, Kispaticeva 12, 10000 Zagreb, Croatia. Tel: +385-91-5624189; fax: +385-1-2367087; e-mail: [email protected] (Z. Colak). Received 2 September 2013; received in revised form 28 February 2014; accepted 1 April 2014
Abstract OBJECTIVES: Postoperative cognitive decline is common after cardiac surgery, but it is often unrecognized at the time of hospital discharge. However, it has a great impact on patient’s quality of life. Cerebral oximetry with the INVOS (IN Vivo Optical Spectroscopy) system provides the possibility of non-invasive, continuous measurement of regional cerebral oxygen saturation (rSO2), which can improve patients’ outcome. The aim of this study was to examine whether cerebral oximetry can decrease the incidence of cognitive decline after coronary artery bypass grafting. METHODS: We have performed a prospective, randomized study with 200 patients enrolled. Patients were divided into INVOS interventional group and CONTROL group without monitoring of cerebral oximetry. A standardized interventional protocol was performed in the INVOS group to maintain rSO2 above 80% of the patient’s baseline value or above 50% of the absolute value. Cognitive evaluation was performed in all patients before and 7 days after surgery. Logistic regression was used to reveal predictors of cognitive decline. RESULTS: The incidence of cognitive decline 7 days after surgery was significantly lower (P = 0.002) in the INVOS interventional group (28%) than in the CONTROL group (52%). Intraoperative use of INVOS monitoring was associated with lower incidence of cognitive decline (odds ratio 0.21). In addition, predictors of cognitive decline revealed by multivariate logistic regression were older age, higher EuroSCORE and SAPS II (Simplified Acute Physiology Score) values, lower educational level and persistence of preoperative atrial fibrillation. Patients with prolonged rSO2 desaturation, defined as rSO2 area under the curve (AUC) of more than 150 min% for desaturation below 20% of baseline or AUC of more than 50 min% for desaturation below 50% absolute value, had an increased risk of cognitive decline. CONCLUSION: Postoperative cognitive outcome was significantly better in patients with intraoperative cerebral oximetry monitoring. Prolonged rSO2 desaturation is a predictor of cognitive decline and has to be avoided. Keywords: Cerebral oximetry • Cardiopulmonary bypass • Cognitive dysfunction
INTRODUCTION Cerebral injury following cardiac surgery is still frequent and has detrimental and often long-lasting consequences. It occurs in various forms, ranging from stroke, which is the most serious but relatively rare, to more frequent cognitive dysfunction (more than 50%), which may not be life threatening but has significant impact on the patient’s quality of life following surgery [1, 2]. The aetiology of these complications is most likely multifactorial, involving embolism, hypoperfusion, hypoxia and inflammatory response, †
Results of the study were presented at a meeting ‘Hot topics in cardiac anesthesia and intensive care’, Bol, Island of Brac, Croatia on 8 September 2012. Results are part of the thesis titled ‘Cognitive dysfunction and the role of cerebral oximetry in patients with coronary artery bypass graft surgery’, 2011. ‡
which eventually lead to regional or global imbalance between cerebral oxygen demand and supply [3–9]. A large number of reports have revealed that the monitoring of regional cerebral oxygen saturation (rSO2) with an INVOS (IN Vivo Optical Spectroscopy) system provides an opportunity for early recognition of this imbalance and subsequent performance of standardized interventions to prevent prolonged rSO2 desaturation with the aim to avoid neurological and other major complications [10–13]. The INVOS system is based on transcranial near-infrared spectroscopy (NIRS), which provides continuous non-invasive measurement of rSO2 in a small local area of cortex [14]. We present the results of the prospective, randomized study of 200 coronary artery bypass grafting (CABG) patients with the use of cardiopulmonary bypass (CPB). Patients were randomized into
© The Author 2014. Published by Oxford University Press on behalf of the European Association for Cardio-Thoracic Surgery. All rights reserved.
ADULT CARDIAC
ORIGINAL ARTICLE
European Journal of Cardio-Thoracic Surgery (2014) 1–8 doi:10.1093/ejcts/ezu193
2
Z. Colak et al. / European Journal of Cardio-Thoracic Surgery
the interventional INVOS group, in which rSO2 was maintained above 80% of patient’s baseline value during the operation, and the CONTROL group, in which monitoring with the INVOS system was not used. The main objectives of the study were to examine whether the patients monitored with the INVOS system had lower incidence of cognitive dysfunction compared with those who were not monitored and to reveal predictive factors of cognitive decline after CABG surgery.
of cerebral oxygenation were involved neither in data collection nor in cognitive test results of patients. Investigators who performed cognitive tests were blinded to patient’s allocation. Patients were also blinded to the allocation group. During surgery, patients involved in the INVOS group were monitored by cerebral oximetry using the INVOS system. Patients in the CONTROL group were not monitored with cerebral oximetry and only standardized monitoring in cardiac surgery was utilized. Figure 1 shows a flow diagram of patients through each stage of a randomized trial.
MATERIALS AND METHODS Study population The data were collected at the University Hospital Center Zagreb, and the study was approved by the hospital ethical committee. The study was preregistered in the ClinicalTrials.gov public database, and the registry number is NCT00917124. Two hundred patients between 40 and 80 years who underwent on-pump CABG surgery and signed informed consent were included in the study. These patients were recruited between June 2009 and September 2010. The study was designed as randomized, prospective and double blinded. Exclusion criteria were significant carotid artery stenosis, previous stroke or head injury, seizure, psychiatric illness, decompensated heart failure (NYHA III/IV), left ventricular ejection fraction less than 25%, emergency cardiac surgery, off-pump CABG and severely impaired renal and liver function.
Operation protocol All patients received the same anaesthesia protocol, including premedication with intramuscular morphine 0.1 mg/kg and induction with midazolam, hypnomidate, sufentanyl and pancuronium. Anaesthesia was maintained with continuous sufentanyl infusion, propofol infusion (during CPB) and sevoflurane. Haemodynamic measurements were performed using the thermodilution method by a Swan Ganz catheter. All patients had standard intraoperative monitoring. Bispectral index or any other neurological monitoring were not used. After heparinization, mild hypothermic (32°C) CPB was established with flow rates of 2–2.5 l/min and haematocrit was maintained above 22%, mean arterial pressure over 60 mmHg and partial pressure of carbon dioxide between 4.5 and 6 kPa by alpha-stat management. There was no possibility to measure haematocrit values by optical sensor incorporated in CPB circuit during CPB. Therefore, they were obtained by analysing blood samples using a point-of-care analyser. Generally, a single-clamp technique was utilized. The rewarming rate during CPB was slow and it took approximately half an hour, without the occurrence of hyperthermia.
Cerebral oximetry All patients in the INVOS group were monitored with the INVOS system (INVOS 5100C; Somanetics Corp., Troy, MI, USA). The INVOS system uses NIRS for non-invasive and continuous measurement of changes in cerebral oxygen saturation. The probes for INVOS cerebral oxygen monitoring were attached bilaterally on the patient’s forehead overlying the frontal–temporal region to awaken patients who breathed oxygen by nasal catheter, just before induction of anaesthesia. A baseline regional cerebral oxygen saturation (rSO2) value for each side of the brain was determined a short time after probes were attached. The rSO2 values were displayed on a screen and recorded during the entire surgery. If the rSO2 during surgery decreased below 80% of baseline value or below 50% of absolute value, we responded with standardized interventions to maintain rSO2 above those values [14]. These interventions involved measures to eliminate mechanical obstruction to cerebral flow (repositioning of head or bypass cannulae), to increase cerebral oxygen delivery (increasing FiO2, pCO2, mean arterial pressure, cardiac output or pump flow and haematocrit) or to reduce cerebral oxygen consumption (increasing of anaesthetic depth and reduction of temperature). Despite those interventions, an inability of maintaining rSO2 above the threshold of 20% under the baseline value was observed in some patients. Prolonged desaturation was determined as the rSO2 area under the curve (AUC) of more than 150 min% under 20% of baseline value, or rSO2 AUC of more than 50 min% under 50% of absolute value according to data from earlier studies [12, 13]. After surgery, AUCs were calculated for desaturations: more than 20% of baseline value (rSO2 AUC < 20% of baseline value), more than 50% (rSO2 AUC < 50% absolute value) and more than 40% of absolute value (rSO2 AUC < 40% absolute value). AUCs account for both depth and duration of desaturation below these thresholds so they are expressed as min% values. If values of AUC from the left and right sides of the brain were different, we took the greater AUC value as relevant.
Neurocognitive assessment—primary outcome measure
Randomization protocol All patients were randomly assigned into the INVOS group or the CONTROL group using a computerized random number generator. After informed consent was obtained, an enclosed assignment in a sequentially numbered, opaque, sealed envelope was allocated to each patient. This envelope was opened by anaesthesiologist at the theatre, and the card inside determined whether the patient was in the INVOS or in the CONTROL group. The anaesthesiologists who performed anaesthesia and interventional protocol for maintaining
Neurocognitive evaluation was performed preoperatively and 7 days after surgery using standardized tests [Mini-Mental State Examination (MMSE), Color Trail Test 1 (CTT 1) and GroovedPegboard Test (GP test)]. Cognitive impairment was defined as a decline in postoperative performance in one or more tests: decrease in an MMSE score for three or more points from baseline value and decrease of 1 SD (standard deviation) or more in performance on CTT 1 and GP tests. According to this definition, the cognitive outcome was assessed for each patient (dichotomous
3
ADULT CARDIAC
Z. Colak et al. / European Journal of Cardio-Thoracic Surgery
Figure 1: Flow diagram of patients allocated in the study.
outcome). Furthermore, all patients were divided into two groups according to the presence or absence of cognitive decline, and statistical analysis of preoperative, intraoperative and postoperative data was performed.
Secondary outcome measures Patients were assessed for neurological deficit presented as coma, stupor, transient ischaemic attack (TIA) or stroke and for the second group of neuropsychological deficits presented as delirium. Coma is defined as a profound state of unconsciousness without response to verbal call, pain or any other stimulus. Stupor is defined as a state of unconsciousness from which the patient can be awakened only by vigorous physical stimulation. Stroke is characterized by the acute onset of a focal neurological deficit lasting more than 24 h. TIA is a clinical syndrome characterized by an acute loss of focal cerebral function lasting less than 24 h. Delirium or encephalopathy is characterized by confusion, agitation, disorientation and decreased alertness, but without an obvious focal neurological deficit. Patients were also evaluated for major organ complications and for duration of intensive care unit (ICU) and overall hospital stay. Perioperative myocardial infarction was defined as an increase
in troponin T values >10 × 99th percentile URL (upper reference limit) during the first 48 h following CABG, when associated with the appearance of new pathological Q-waves or new left bundle branch blocks. All patients were on telemetry until hospital discharge and episodes of atrial fibrillation were observed. The duration of mechanical ventilation was defined as a period from the arrival in ICU until extubation. Prolonged mechanical ventilation was defined as the need for mechanical ventilation longer than 24 h. Infection was defined as clinically manifested and microbiologically confirmed infection. A logistic value of EuroSCORE was calculated for all patients. SAPS II (Simplified Acute Physiology Score) is a severity of disease classification system for patients admitted to ICUs. SAPS II values were calculated at admission in ICU and at discharge from ICU. Carotid arteries Doppler ultrasound was routinely undertaken in all patients older than 65 years and in diabetic patients older than 60 years.
Statistical analysis Sample size was determined assuming 50% incidence of cognitive impairment after cardiac surgery and the possibility of decreasing that incidence to 30% using cerebral oximetry. Based on 0.8
Z. Colak et al. / European Journal of Cardio-Thoracic Surgery
4
power to detect a significant difference (P = 0.05), 90 patients were required for each study group. We enrolled 100 patients per group due to the potential loss in the follow-up. All statistical analyses were performed using Statistica 7.1 and SAS 8.2 software. Categorical data were analysed using the χ 2 test or Fisher’s exact test, which is used when the expected frequency in any cell of a contingency table was below 5. According to normality of distribution, continuous variables were compared between groups using Student’s t-test or non-parametric Mann–Whitney U-test as appropriate. Multivariate logistic regression analysis with a stepwise procedure was used to define the risk factors for the development of cognitive decline. Two multivariate logistic models were made separately for preoperative and operative variables, listed as an appendix to Tables 3 and 4. The entrance level into multivariate
logistic regression analysis was set to a P–value of less than or equal to 0.15. Univariate logistic regression analysis was also performed for all variables. Continuous data are presented as mean ± SD. Categorical variables are presented as percentages. Statistical significance was considered at P-value of less than 0.05.
RESULTS Figure 1 shows the flow diagram with information about enrolment, intervention allocation, follow-up and analysis of patients. There was good comparability between the two studied groups (Table 1). No statistically significant differences were observed between the INVOS and CONTROL group in patients’
Table 1: Comparison between the INVOS and CONTROL groups regarding demographic and clinical data Variables
Age (years) Age categorical (% of patients ≥65 years) Gender (female %) EuroSCORE (%) Left ventricular ejection fraction (%) Years of education Comorbidity (% of patients) Diabetes Hypertension Myocardial infarction within 1 month Atrial fibrillation Renal insufficiency Number of grafts Bypass time (min) Aortic cross-clamp time (min) Cardiac index before CPB (l/min/m2) Cardiac index after CPB (l/min/m2) ‘Single-clamp’ technique (% of patients) Erythrocytes concentrate during CPB (ml) Erythrocytes concentrate after CPB (ml) W/o Erythrocytes concentrate during surgery (% of patients) W/o Erythrocytes concentrate during hospitalization (% of patients) Inotropics (% of patients) Erythrocytes concentrate in ICU (ml) Cardiac index—ICU arrival (l/min/m2) Cardiac index—after 24 h (l/min/m2) Mechanical ventilation (h) Postoperative complications (% of patients) Stroke, TIA, coma, stupor Delirium Cognitive impairment Myocardial infarction Atrial fibrillation Prolonged mechanical ventilation Haemodialysis Infection Revision ICU length of stay (days) Hospital stay >7 days (% of patients) SAPS II—arrival in ICU SAPS II—ICU discharge
P-value
Group INVOS (n = 94)
Control (n = 96)
61.9 ± 7.1 32 20 2.2 ± 1.7 56 ± 9.7 11.3 ± 3.4
63.4 ± 8.8 50 24 2.4 ± 1.7 56 ± 9.9 11.2 ± 3.2
0.177 0.017a 0.655 0.282 0.765 0.811
30 84 9 4 2 2.5 ± 0.6 91 ± 31 63 ± 23 2.3 ± 0.4 2.9 ± 0.6 86 265 ± 306 130 ± 212 36 18 46 234 ± 333 2.9 ± 0.5 3.1 ± 0.5 9.7 ± 4.2
34 88 13 5 1 2.6 ± 0.6 89 ± 32 62 ± 26 2.3 ± 0.5 2.9 ± 0.6 79 181 ± 283 214 ± 294 36 24 40 239 ± 317 2.9 ± 0.6 3.1 ± 0.5 9.3 ± 4.1
0.602 0.367 0.370 0.757 0.619 0.640 0.596 0.669 0.722 0.928 0.203 0.056 0.055 0.967 0.415 0.391 0.913 0.461 0.648 0.551
4 9 28 8 23 1 0 18 0 2.7 ± 6.2 49 18.9 ± 5.5 11.1 ± 3.6
1 14 52 7 27 2 1 19 1 1.9 ± 0.9 45 19.2 ± 5.5 11.4 ± 4.0
0.204 0.292 0.002a 0.934 0.500 1.000 1.000 0.962 1.000 0.443 0.729 0.655 0.565
Data are expressed as frequencies % or X ± SD (mean ± standard deviation). CPB: cardiopulmonary bypass; ICU: intensive care unit; SAPS: Simplified Acute Physiology Score; TIA: transient ischaemic attack; INVOS: IN Vivo Optical Spectroscopy. a Statistically significant value.
Z. Colak et al. / European Journal of Cardio-Thoracic Surgery
The results from multivariate logistic regression of all patients revealed that intraoperative use of INVOS monitoring was associated with lower incidence of cognitive decline [odds ratio (OR) 0.21] (Table 3). In addition, predictors of postoperative cognitive decline were older age, lower educational level, higher EuroSCORE value, preoperative persistence of atrial fibrillation and higher value of SAPS II at the moment of ICU discharge. The results from univariate logistic regression analysis are also presented in Table 3. Multivariate analysis within the INVOS group revealed prolonged rSO2 desaturation during surgery as an important risk factor for cognitive decline (Table 4).
DISCUSSION In the present study, the overall incidence of early postoperative cognitive impairment was 40.3%, defined as a decrease in
Table 2: Comparison between patients with and without cognitive decline Variables
Without cognitive decline (n = 108)
With cognitive decline (n = 73)
P-value
Age (years) Gender (female %) EuroSCORE (%) Left ventricular ejection fraction (%) Years of education Comorbidity (% of patients) Diabetes Hypertension Myocardial infarction within 1 month Atrial fibrillation Renal insufficiency Number of grafts Bypass time (min) Aortic cross-clamp time (min) Cardiac index before CPB (l/min/m2) Cardiac index after CPB (l/min/m2) ‘Single-clamp’ technique (% of patients) Erythrocytes concentrate during CPB (ml) Erythrocytes concentrate after CPB (ml) W/o Erythrocytes concentrate during surgery (% of patients) W/o Erythrocytes concentrate during hospitalization (% of patients) Inotropics (% of patients) Erythrocytes concentrate in ICU (ml) Cardiac index—ICU arrival (l/min/m2) Cardiac index—after 24 h (l/min/m2) Mechanical ventilation (h) Postoperative complications (% of patients) Stroke, TIA, coma, stupor Delirium Myocardial infarction Atrial fibrillation Prolonged mechanical ventilation Haemodialysis Infection Revision ICU length of stay (days) Hospital stay >7 days (% of patients) SAPS II (arrival in ICU) SAPS II (ICU discharge)
59.6 ± 7.4 26 1.9 ± 1.4 56.1 ± 9.5 11.8 ± 2.9
66.9 ± 6.9 21 2.9 ± 1.9 56.2 ± 9.8 10.6 ± 3.7
Influence of intraoperative cerebral oximetry monitoring on neurocognitive function after coronary artery bypass surgery: a randomized, prospective study†‡ Zeljko Colaka,*, Marko Borojevicb, Anamarija Bogovicc, Visnja Ivancana, Bojan Biocinab and Visnja Majeric-Koglera a b c
Department of Anesthesiology, Reanimatology and Intensive Care Medicine, University Hospital Center Zagreb, Zagreb, Croatia Department of Cardiac Surgery, University Hospital Center Zagreb, Zagreb, Croatia Psychiatric Hospital Sv. Ivan, Zagreb, Croatia
* Corresponding author. Department of Anesthesiology, Reanimatology and Intensive Care Medicine, University Hospital Center Zagreb, Kispaticeva 12, 10000 Zagreb, Croatia. Tel: +385-91-5624189; fax: +385-1-2367087; e-mail: [email protected] (Z. Colak). Received 2 September 2013; received in revised form 28 February 2014; accepted 1 April 2014
Abstract OBJECTIVES: Postoperative cognitive decline is common after cardiac surgery, but it is often unrecognized at the time of hospital discharge. However, it has a great impact on patient’s quality of life. Cerebral oximetry with the INVOS (IN Vivo Optical Spectroscopy) system provides the possibility of non-invasive, continuous measurement of regional cerebral oxygen saturation (rSO2), which can improve patients’ outcome. The aim of this study was to examine whether cerebral oximetry can decrease the incidence of cognitive decline after coronary artery bypass grafting. METHODS: We have performed a prospective, randomized study with 200 patients enrolled. Patients were divided into INVOS interventional group and CONTROL group without monitoring of cerebral oximetry. A standardized interventional protocol was performed in the INVOS group to maintain rSO2 above 80% of the patient’s baseline value or above 50% of the absolute value. Cognitive evaluation was performed in all patients before and 7 days after surgery. Logistic regression was used to reveal predictors of cognitive decline. RESULTS: The incidence of cognitive decline 7 days after surgery was significantly lower (P = 0.002) in the INVOS interventional group (28%) than in the CONTROL group (52%). Intraoperative use of INVOS monitoring was associated with lower incidence of cognitive decline (odds ratio 0.21). In addition, predictors of cognitive decline revealed by multivariate logistic regression were older age, higher EuroSCORE and SAPS II (Simplified Acute Physiology Score) values, lower educational level and persistence of preoperative atrial fibrillation. Patients with prolonged rSO2 desaturation, defined as rSO2 area under the curve (AUC) of more than 150 min% for desaturation below 20% of baseline or AUC of more than 50 min% for desaturation below 50% absolute value, had an increased risk of cognitive decline. CONCLUSION: Postoperative cognitive outcome was significantly better in patients with intraoperative cerebral oximetry monitoring. Prolonged rSO2 desaturation is a predictor of cognitive decline and has to be avoided. Keywords: Cerebral oximetry • Cardiopulmonary bypass • Cognitive dysfunction
INTRODUCTION Cerebral injury following cardiac surgery is still frequent and has detrimental and often long-lasting consequences. It occurs in various forms, ranging from stroke, which is the most serious but relatively rare, to more frequent cognitive dysfunction (more than 50%), which may not be life threatening but has significant impact on the patient’s quality of life following surgery [1, 2]. The aetiology of these complications is most likely multifactorial, involving embolism, hypoperfusion, hypoxia and inflammatory response, †
Results of the study were presented at a meeting ‘Hot topics in cardiac anesthesia and intensive care’, Bol, Island of Brac, Croatia on 8 September 2012. Results are part of the thesis titled ‘Cognitive dysfunction and the role of cerebral oximetry in patients with coronary artery bypass graft surgery’, 2011. ‡
which eventually lead to regional or global imbalance between cerebral oxygen demand and supply [3–9]. A large number of reports have revealed that the monitoring of regional cerebral oxygen saturation (rSO2) with an INVOS (IN Vivo Optical Spectroscopy) system provides an opportunity for early recognition of this imbalance and subsequent performance of standardized interventions to prevent prolonged rSO2 desaturation with the aim to avoid neurological and other major complications [10–13]. The INVOS system is based on transcranial near-infrared spectroscopy (NIRS), which provides continuous non-invasive measurement of rSO2 in a small local area of cortex [14]. We present the results of the prospective, randomized study of 200 coronary artery bypass grafting (CABG) patients with the use of cardiopulmonary bypass (CPB). Patients were randomized into
© The Author 2014. Published by Oxford University Press on behalf of the European Association for Cardio-Thoracic Surgery. All rights reserved.
ADULT CARDIAC
ORIGINAL ARTICLE
European Journal of Cardio-Thoracic Surgery (2014) 1–8 doi:10.1093/ejcts/ezu193
2
Z. Colak et al. / European Journal of Cardio-Thoracic Surgery
the interventional INVOS group, in which rSO2 was maintained above 80% of patient’s baseline value during the operation, and the CONTROL group, in which monitoring with the INVOS system was not used. The main objectives of the study were to examine whether the patients monitored with the INVOS system had lower incidence of cognitive dysfunction compared with those who were not monitored and to reveal predictive factors of cognitive decline after CABG surgery.
of cerebral oxygenation were involved neither in data collection nor in cognitive test results of patients. Investigators who performed cognitive tests were blinded to patient’s allocation. Patients were also blinded to the allocation group. During surgery, patients involved in the INVOS group were monitored by cerebral oximetry using the INVOS system. Patients in the CONTROL group were not monitored with cerebral oximetry and only standardized monitoring in cardiac surgery was utilized. Figure 1 shows a flow diagram of patients through each stage of a randomized trial.
MATERIALS AND METHODS Study population The data were collected at the University Hospital Center Zagreb, and the study was approved by the hospital ethical committee. The study was preregistered in the ClinicalTrials.gov public database, and the registry number is NCT00917124. Two hundred patients between 40 and 80 years who underwent on-pump CABG surgery and signed informed consent were included in the study. These patients were recruited between June 2009 and September 2010. The study was designed as randomized, prospective and double blinded. Exclusion criteria were significant carotid artery stenosis, previous stroke or head injury, seizure, psychiatric illness, decompensated heart failure (NYHA III/IV), left ventricular ejection fraction less than 25%, emergency cardiac surgery, off-pump CABG and severely impaired renal and liver function.
Operation protocol All patients received the same anaesthesia protocol, including premedication with intramuscular morphine 0.1 mg/kg and induction with midazolam, hypnomidate, sufentanyl and pancuronium. Anaesthesia was maintained with continuous sufentanyl infusion, propofol infusion (during CPB) and sevoflurane. Haemodynamic measurements were performed using the thermodilution method by a Swan Ganz catheter. All patients had standard intraoperative monitoring. Bispectral index or any other neurological monitoring were not used. After heparinization, mild hypothermic (32°C) CPB was established with flow rates of 2–2.5 l/min and haematocrit was maintained above 22%, mean arterial pressure over 60 mmHg and partial pressure of carbon dioxide between 4.5 and 6 kPa by alpha-stat management. There was no possibility to measure haematocrit values by optical sensor incorporated in CPB circuit during CPB. Therefore, they were obtained by analysing blood samples using a point-of-care analyser. Generally, a single-clamp technique was utilized. The rewarming rate during CPB was slow and it took approximately half an hour, without the occurrence of hyperthermia.
Cerebral oximetry All patients in the INVOS group were monitored with the INVOS system (INVOS 5100C; Somanetics Corp., Troy, MI, USA). The INVOS system uses NIRS for non-invasive and continuous measurement of changes in cerebral oxygen saturation. The probes for INVOS cerebral oxygen monitoring were attached bilaterally on the patient’s forehead overlying the frontal–temporal region to awaken patients who breathed oxygen by nasal catheter, just before induction of anaesthesia. A baseline regional cerebral oxygen saturation (rSO2) value for each side of the brain was determined a short time after probes were attached. The rSO2 values were displayed on a screen and recorded during the entire surgery. If the rSO2 during surgery decreased below 80% of baseline value or below 50% of absolute value, we responded with standardized interventions to maintain rSO2 above those values [14]. These interventions involved measures to eliminate mechanical obstruction to cerebral flow (repositioning of head or bypass cannulae), to increase cerebral oxygen delivery (increasing FiO2, pCO2, mean arterial pressure, cardiac output or pump flow and haematocrit) or to reduce cerebral oxygen consumption (increasing of anaesthetic depth and reduction of temperature). Despite those interventions, an inability of maintaining rSO2 above the threshold of 20% under the baseline value was observed in some patients. Prolonged desaturation was determined as the rSO2 area under the curve (AUC) of more than 150 min% under 20% of baseline value, or rSO2 AUC of more than 50 min% under 50% of absolute value according to data from earlier studies [12, 13]. After surgery, AUCs were calculated for desaturations: more than 20% of baseline value (rSO2 AUC < 20% of baseline value), more than 50% (rSO2 AUC < 50% absolute value) and more than 40% of absolute value (rSO2 AUC < 40% absolute value). AUCs account for both depth and duration of desaturation below these thresholds so they are expressed as min% values. If values of AUC from the left and right sides of the brain were different, we took the greater AUC value as relevant.
Neurocognitive assessment—primary outcome measure
Randomization protocol All patients were randomly assigned into the INVOS group or the CONTROL group using a computerized random number generator. After informed consent was obtained, an enclosed assignment in a sequentially numbered, opaque, sealed envelope was allocated to each patient. This envelope was opened by anaesthesiologist at the theatre, and the card inside determined whether the patient was in the INVOS or in the CONTROL group. The anaesthesiologists who performed anaesthesia and interventional protocol for maintaining
Neurocognitive evaluation was performed preoperatively and 7 days after surgery using standardized tests [Mini-Mental State Examination (MMSE), Color Trail Test 1 (CTT 1) and GroovedPegboard Test (GP test)]. Cognitive impairment was defined as a decline in postoperative performance in one or more tests: decrease in an MMSE score for three or more points from baseline value and decrease of 1 SD (standard deviation) or more in performance on CTT 1 and GP tests. According to this definition, the cognitive outcome was assessed for each patient (dichotomous
3
ADULT CARDIAC
Z. Colak et al. / European Journal of Cardio-Thoracic Surgery
Figure 1: Flow diagram of patients allocated in the study.
outcome). Furthermore, all patients were divided into two groups according to the presence or absence of cognitive decline, and statistical analysis of preoperative, intraoperative and postoperative data was performed.
Secondary outcome measures Patients were assessed for neurological deficit presented as coma, stupor, transient ischaemic attack (TIA) or stroke and for the second group of neuropsychological deficits presented as delirium. Coma is defined as a profound state of unconsciousness without response to verbal call, pain or any other stimulus. Stupor is defined as a state of unconsciousness from which the patient can be awakened only by vigorous physical stimulation. Stroke is characterized by the acute onset of a focal neurological deficit lasting more than 24 h. TIA is a clinical syndrome characterized by an acute loss of focal cerebral function lasting less than 24 h. Delirium or encephalopathy is characterized by confusion, agitation, disorientation and decreased alertness, but without an obvious focal neurological deficit. Patients were also evaluated for major organ complications and for duration of intensive care unit (ICU) and overall hospital stay. Perioperative myocardial infarction was defined as an increase
in troponin T values >10 × 99th percentile URL (upper reference limit) during the first 48 h following CABG, when associated with the appearance of new pathological Q-waves or new left bundle branch blocks. All patients were on telemetry until hospital discharge and episodes of atrial fibrillation were observed. The duration of mechanical ventilation was defined as a period from the arrival in ICU until extubation. Prolonged mechanical ventilation was defined as the need for mechanical ventilation longer than 24 h. Infection was defined as clinically manifested and microbiologically confirmed infection. A logistic value of EuroSCORE was calculated for all patients. SAPS II (Simplified Acute Physiology Score) is a severity of disease classification system for patients admitted to ICUs. SAPS II values were calculated at admission in ICU and at discharge from ICU. Carotid arteries Doppler ultrasound was routinely undertaken in all patients older than 65 years and in diabetic patients older than 60 years.
Statistical analysis Sample size was determined assuming 50% incidence of cognitive impairment after cardiac surgery and the possibility of decreasing that incidence to 30% using cerebral oximetry. Based on 0.8
Z. Colak et al. / European Journal of Cardio-Thoracic Surgery
4
power to detect a significant difference (P = 0.05), 90 patients were required for each study group. We enrolled 100 patients per group due to the potential loss in the follow-up. All statistical analyses were performed using Statistica 7.1 and SAS 8.2 software. Categorical data were analysed using the χ 2 test or Fisher’s exact test, which is used when the expected frequency in any cell of a contingency table was below 5. According to normality of distribution, continuous variables were compared between groups using Student’s t-test or non-parametric Mann–Whitney U-test as appropriate. Multivariate logistic regression analysis with a stepwise procedure was used to define the risk factors for the development of cognitive decline. Two multivariate logistic models were made separately for preoperative and operative variables, listed as an appendix to Tables 3 and 4. The entrance level into multivariate
logistic regression analysis was set to a P–value of less than or equal to 0.15. Univariate logistic regression analysis was also performed for all variables. Continuous data are presented as mean ± SD. Categorical variables are presented as percentages. Statistical significance was considered at P-value of less than 0.05.
RESULTS Figure 1 shows the flow diagram with information about enrolment, intervention allocation, follow-up and analysis of patients. There was good comparability between the two studied groups (Table 1). No statistically significant differences were observed between the INVOS and CONTROL group in patients’
Table 1: Comparison between the INVOS and CONTROL groups regarding demographic and clinical data Variables
Age (years) Age categorical (% of patients ≥65 years) Gender (female %) EuroSCORE (%) Left ventricular ejection fraction (%) Years of education Comorbidity (% of patients) Diabetes Hypertension Myocardial infarction within 1 month Atrial fibrillation Renal insufficiency Number of grafts Bypass time (min) Aortic cross-clamp time (min) Cardiac index before CPB (l/min/m2) Cardiac index after CPB (l/min/m2) ‘Single-clamp’ technique (% of patients) Erythrocytes concentrate during CPB (ml) Erythrocytes concentrate after CPB (ml) W/o Erythrocytes concentrate during surgery (% of patients) W/o Erythrocytes concentrate during hospitalization (% of patients) Inotropics (% of patients) Erythrocytes concentrate in ICU (ml) Cardiac index—ICU arrival (l/min/m2) Cardiac index—after 24 h (l/min/m2) Mechanical ventilation (h) Postoperative complications (% of patients) Stroke, TIA, coma, stupor Delirium Cognitive impairment Myocardial infarction Atrial fibrillation Prolonged mechanical ventilation Haemodialysis Infection Revision ICU length of stay (days) Hospital stay >7 days (% of patients) SAPS II—arrival in ICU SAPS II—ICU discharge
P-value
Group INVOS (n = 94)
Control (n = 96)
61.9 ± 7.1 32 20 2.2 ± 1.7 56 ± 9.7 11.3 ± 3.4
63.4 ± 8.8 50 24 2.4 ± 1.7 56 ± 9.9 11.2 ± 3.2
0.177 0.017a 0.655 0.282 0.765 0.811
30 84 9 4 2 2.5 ± 0.6 91 ± 31 63 ± 23 2.3 ± 0.4 2.9 ± 0.6 86 265 ± 306 130 ± 212 36 18 46 234 ± 333 2.9 ± 0.5 3.1 ± 0.5 9.7 ± 4.2
34 88 13 5 1 2.6 ± 0.6 89 ± 32 62 ± 26 2.3 ± 0.5 2.9 ± 0.6 79 181 ± 283 214 ± 294 36 24 40 239 ± 317 2.9 ± 0.6 3.1 ± 0.5 9.3 ± 4.1
0.602 0.367 0.370 0.757 0.619 0.640 0.596 0.669 0.722 0.928 0.203 0.056 0.055 0.967 0.415 0.391 0.913 0.461 0.648 0.551
4 9 28 8 23 1 0 18 0 2.7 ± 6.2 49 18.9 ± 5.5 11.1 ± 3.6
1 14 52 7 27 2 1 19 1 1.9 ± 0.9 45 19.2 ± 5.5 11.4 ± 4.0
0.204 0.292 0.002a 0.934 0.500 1.000 1.000 0.962 1.000 0.443 0.729 0.655 0.565
Data are expressed as frequencies % or X ± SD (mean ± standard deviation). CPB: cardiopulmonary bypass; ICU: intensive care unit; SAPS: Simplified Acute Physiology Score; TIA: transient ischaemic attack; INVOS: IN Vivo Optical Spectroscopy. a Statistically significant value.
Z. Colak et al. / European Journal of Cardio-Thoracic Surgery
The results from multivariate logistic regression of all patients revealed that intraoperative use of INVOS monitoring was associated with lower incidence of cognitive decline [odds ratio (OR) 0.21] (Table 3). In addition, predictors of postoperative cognitive decline were older age, lower educational level, higher EuroSCORE value, preoperative persistence of atrial fibrillation and higher value of SAPS II at the moment of ICU discharge. The results from univariate logistic regression analysis are also presented in Table 3. Multivariate analysis within the INVOS group revealed prolonged rSO2 desaturation during surgery as an important risk factor for cognitive decline (Table 4).
DISCUSSION In the present study, the overall incidence of early postoperative cognitive impairment was 40.3%, defined as a decrease in
Table 2: Comparison between patients with and without cognitive decline Variables
Without cognitive decline (n = 108)
With cognitive decline (n = 73)
P-value
Age (years) Gender (female %) EuroSCORE (%) Left ventricular ejection fraction (%) Years of education Comorbidity (% of patients) Diabetes Hypertension Myocardial infarction within 1 month Atrial fibrillation Renal insufficiency Number of grafts Bypass time (min) Aortic cross-clamp time (min) Cardiac index before CPB (l/min/m2) Cardiac index after CPB (l/min/m2) ‘Single-clamp’ technique (% of patients) Erythrocytes concentrate during CPB (ml) Erythrocytes concentrate after CPB (ml) W/o Erythrocytes concentrate during surgery (% of patients) W/o Erythrocytes concentrate during hospitalization (% of patients) Inotropics (% of patients) Erythrocytes concentrate in ICU (ml) Cardiac index—ICU arrival (l/min/m2) Cardiac index—after 24 h (l/min/m2) Mechanical ventilation (h) Postoperative complications (% of patients) Stroke, TIA, coma, stupor Delirium Myocardial infarction Atrial fibrillation Prolonged mechanical ventilation Haemodialysis Infection Revision ICU length of stay (days) Hospital stay >7 days (% of patients) SAPS II (arrival in ICU) SAPS II (ICU discharge)
59.6 ± 7.4 26 1.9 ± 1.4 56.1 ± 9.5 11.8 ± 2.9
66.9 ± 6.9 21 2.9 ± 1.9 56.2 ± 9.8 10.6 ± 3.7
